AI Pioneer Pushes for US-Led Global Watchdog to Police Frontier Models

Demis Hassabis, the CEO and co-founder of Google DeepMind, is calling for the creation of a global AI oversight body - and he wants the United States to run it. In a new blog post, Hassabis argues that as artificial intelligence systems become more powerful, the world needs a regulator with real teeth: an organization that can review cutting-edge models before they go public and, if necessary, order an industry-wide pause on deployment. He envisions a body modeled after existing financial regulators like FINRA, staffed by independent experts and representatives from open-source communities.

Hassabis, who shared the 2024 Nobel Prize in Chemistry for his work on AI-driven protein prediction, says the window for action is narrowing. He warns that artificial general intelligence - machines that can think and learn like humans - is likely just a few years away. In his post, titled "A Framework for Frontier AI and the Dawning of a New Age," he describes the current moment as humanity standing "in the foothills of the singularity." According to Axios, Hassabis has spent months quietly building support for the proposal, briefing the Trump administration, other AI labs, and European officials. He told Axios the response from the White House has been "very positive."

The proposal comes amid a broader push from industry leaders for coherent AI governance. There is currently no global set of rules for AI, and the U.S. lacks a comprehensive national framework. Hassabis recently signed a statement calling for stronger safeguards against AI-enabled bioweapons. His latest comments follow a separate warning from top economists and tech figures - including Anthropic cofounder Jack Clark and former Google CEO Eric Schmidt - who urged world leaders to prepare for AI's economic fallout. Hassabis hopes to have his proposed watchdog up and running before the end of the year.

Source: The Verge